JPH10271049A - Echo canceller double talk monitor - Google Patents

Abstract

(57) [Problem] To provide a double talk monitoring device of an echo canceller for reliably detecting occurrence of a double talk state during training. A training signal output unit that outputs a training signal of a certain level, an echo amount calculation unit that calculates an echo amount to be removed based on the training signal output from the training signal output unit, and an echo amount An echo removing unit 4 for removing an echo component in the signal line based on the echo amount calculated by the calculating unit 4, a level detecting unit 6 for detecting a peak level of an input signal to the echo amount calculating unit 4, and a level detecting unit A comparison unit for comparing the level value of the peak level detected by the comparison unit with a preset reference value, and when the level value of the input signal exceeds the reference value as a result of the comparison by the comparison unit, an echo amount calculation unit In step 4, the echo amount is calculated again.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a double talk monitoring apparatus for an echo canceller for detecting a double talk state in a communication line.

[0002]

2. Description of the Related Art Conventionally, a transmission signal flowing through a dedicated 2-wire communication line is converted so that it can be transmitted to a 4-wire dedicated communication line, and a transmission signal flowing through a 4-wire dedicated communication line is dedicated to a 2-wire dedicated communication line. As a device for converting data so that it can be transmitted to a communication line, there is a 2-wire to 4-wire conversion circuit, that is, an in-band ringer. The in-band ringer has a function of superimposing a predetermined control signal within a communication band. This gives 1
Various signals such as a 6 Hz dial pulse signal and an on-hook signal indicating an on-hook or off-hook state are temporarily
The signal is replaced with a signal in a voice band of 300 Hz to 3 kHz, and is transmitted to a communication terminal on the other side, or a signal is received in a reverse procedure.

FIG. 4 shows an example of a connection state of a conventional in-band ringer. As shown in FIG.
0 and 20 are connected via a communication line L, respectively, and each in-band ringer 10 (20) is connected to the private branch exchange 1
1 (21). Furthermore, each private branch exchange 1
The communication terminals 12 (22), 13
(23) is connected. Here, the in-band ringer 10 (20) is provided with an echo canceller in order to minimize the echo component generated at the time of 2-wire to 4-wire conversion.

That is, in a long-distance telephone line, a voice may be delayed due to a delay of its own voice through a two-wire to four-wire conversion circuit on the other side and may cause a remarkable trouble in a call. A device that prevents this is an echo canceller.The echo canceller has a memory that stores the impulse response of the reverberation path. The same echo is generated, and the generated echo component is subtracted from the transmission signal to remove the echo component.

FIGS. 5 to 7 are diagrams for explaining the operation procedure in the conventional example. FIG. 5 is a flowchart for explaining a main process, and FIG. 6 is a diagram for explaining an initial setting process. FIG. 7 is a flowchart for explaining the normal operation process.

First, when the power is turned on (step S3)
01), the station B (the right half in FIG. 4) is selected (step S302), and an initialization process described later is executed (step S303). Subsequently, the station A (the left half in FIG. 4) is selected (step S304), and the initial setting process is executed similarly to the station B (step S305), and the initial setting of the stations B and A is completed. Then, a normal operation process is performed (step S306).

[0007] Steps S303 and S3 in FIG.
In the initial setting process at 05, as shown in FIG. 6, the filter coefficients are forcibly reset (step S401),
Next, a DSP (Digital Signal Process)
er) generates a training signal (step S40).
2) Re-converge the filter coefficients inside the echo canceller (step S403). Then, the re-converged filter coefficient is updated as a new filter coefficient (step S404), and it is checked whether or not synchronization has been established (step S405).

If it is determined in step S405 that synchronization has not been established (step S40).
5; N), waiting until synchronization is established, while
If synchronization is established in the determination processing of step S405 (step S405; Y), step S405 in FIG.
Proceed to 304 or step S306.

In the normal operation process in step S306 in FIG. 5, as shown in FIG. 7, first, it is confirmed whether or not the current operation is in an idle state (step S501). If it is not in the idle state in this determination process (step S501; N), an error occurs (step S50).
2). On the other hand, in the determination process of step S501, when it is confirmed that the vehicle is in the idle state (step S50).
1; Y), and then confirm whether a dial call has been made (step S503) and, if the call has been confirmed (step S50).
3; Y), and confirm whether there is a response to the call (step S504).

If there is no outgoing call in the determination process of step S503 (step S503; N), it is confirmed whether or not there is an incoming call (step S505). And
If there is no incoming call in the determination process of step S505 (step S505; N), the process returns to step S503. Then, in the determination processing of step 505, if there is an incoming call (step S505; Y), it is confirmed whether or not there is a response to the incoming call (step S506).

If there is no response in the determination processing in step S504 or S506 (step S504; N, step S506; N), the process returns to step S503. On the other hand, in the determination processing of step S504 or S506, if there is a response (step S504; Y, step S506;
Y), a call is made between the stations A and B (step S507), and the presence or absence of the call termination is confirmed (step S50).
8).

If the call is not yet ended (step S508; N), the process proceeds to step S507 to continue the call. On the other hand, if it is determined in the determination processing of step S508 that the call is to be ended (step S508; Y), step S508 is performed.
The processing from 503 is repeatedly executed. Thus, when the echo canceller cancels the echo component generated in the communication line, the training of the echo canceller is indispensable.

Specifically, as shown in FIG. 8, an echo as shown in FIG. 9 is generated by flowing a reference training signal through a signal line. In the echo canceller,
A pseudo echo that cancels this echo component is generated and applied to the signal line. This removes the echo component. The training of the echo canceller is
For example, 200 msec to 400 m
This is performed in a short time of about msec.

[0014]

However, in such a conventional echo canceller, when a signal other than the training signal is input during the training period,
Since the echo cancellation value is determined based on the input signal, an erroneous filter value is set, and as shown in FIG. 10, there is a problem that the echo component cannot be canceled sufficiently. The state in which the amount of echo cancellation is insufficient is called the double talk state. If the training fails and the double talk state occurs, the call quality is greatly reduced, so it is necessary to accurately detect the double talk state There is.

SUMMARY OF THE INVENTION An object of the present invention is to solve the above-mentioned problems, and an object of the present invention is to provide a double-talk monitoring device of an echo canceller for reliably detecting occurrence of a double-talk state during training.

[0016]

An echo canceller according to the present invention calculates a training signal output section for outputting a training signal of a constant level, and an echo amount to be removed based on the training signal output from the training signal output section. An echo amount calculating section, an echo removing section for removing an echo component in a signal line based on the echo amount calculated by the echo amount calculating section, and a level detection for detecting a peak level value of an input signal to the echo amount calculating section. And a comparing section for comparing the peak level value detected by the level detecting section with a preset reference value, and when the peak level value of the input signal exceeds the reference value as a result of the comparison by the comparing section, an echo The amount calculator is configured to calculate the echo amount again.

In this case, the reference value in the comparing unit is the same as the peak value of the training signal output from the training signal output unit, or the peak value of the training signal output from the training signal output unit and the reference value in the training signal. It is preferable to set to a value between the peak value of the echo component.

[0018]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, the present invention will be described in detail based on one illustrated embodiment. FIG. 1 is a diagram illustrating a main configuration of an echo canceller according to the present embodiment. In FIG. 1, an in-band ringer 1 includes a hybrid transformer 2 provided on a terminal side, a level monitoring unit 3 connected to a signal line on the receiving side of the hybrid transformer 2, and an echo canceller (echo canceller) connected to the hybrid transformer 2. And a signal processing unit (training signal output unit) 4 provided on the line side for processing a received signal from the echo canceller 4 and a transmission signal to the echo canceller 4.

The hybrid transformer 2 branches the transmission signal and the reception signal based on the directionality of the reception signal, and converts the transmission signal output from the signal line on the terminal side into an echo canceller 4 and a signal processing unit. 5, and transmits a received signal input from the line via the signal processing unit 5 and the echo canceller 4 to the terminal. This prevents the transmission signal from leaking to the receiving side and prevents the receiving signal from leaking to the transmitting side.

The level monitoring section 3 has a signal smoothing section (level detecting section) 6 for connecting an input terminal to a reception signal line via the hybrid and the lance 2, and has one input terminal connected to an output terminal side of the signal smoothing section 6. And a comparator 7 for inputting a predetermined reference value to the other input terminal. The signal smoothing unit 6 smoothes the signal contained in the received signal line, converts the signal into a constant level value, and outputs it to the comparison unit 7. The comparison unit 7 outputs the signal output from the signal smoothing unit 6. The level is compared with a reference value, and when the signal level exceeds the reference value, a control signal for resetting the device power is output.

The echo canceller 4 is provided for removing an echo component generated at the time of 2-wire to 4-wire conversion by the hybrid transformer 2, and calculates a filter coefficient which becomes an optimum value during a training operation. To set. The signal processing unit 5 processes signals transmitted and received between the terminal side and the line side,
It also has a function of sending a training signal through a signal line including the echo canceller 4.

Next, an example of operation of the in-band ringer in the above embodiment will be described with reference to FIGS. FIG. 2 is a flowchart for explaining main operation processing in this embodiment, and FIG. 3 is a flowchart for explaining level monitoring detection processing.

First, when the power supply is turned on (step S1)
01), it is checked whether 200 msec has elapsed since the power-on (step S102). If not (step S102; N), the process stands by until the elapse. On the other hand, in the determination process of step S102, 200 ms
If ec has elapsed (step S102; Y), an initial setting process is executed according to the procedure shown in FIG. 6 (step S10).
3) Subsequently, a level monitoring process described later is executed (step S104).

Next, in the level monitoring process of step S104, it is checked whether or not double talk has occurred (step S105). If double talk has not been detected (step S105; N), FIG. The normal operation process is executed according to the procedure shown in FIG.
6). On the other hand, when the occurrence of double talk is detected in the determination processing of step S105 (step S105;
Y), by resetting the power supply, step S10
The processing is executed again from the processing of step 3.

As shown in FIG. 3, the level monitoring and detection process shown in step S104 of FIG. 2 first takes in an echo signal in a signal line and smoothes the echo level (step S201). Then, the comparison unit 7 compares the smoothed echo level with the reference value (step S20).
2) Here, if the output value of the echo level is larger (step S202; Y), the presence of double talk is set as the occurrence of double talk (step S202).
203) If the output value of the echo level is smaller (step S202; N), it is determined that double talk has not occurred, and no double talk is set (step S204), and the process returns to step S105 in FIG. I do.

As described above, in the present embodiment, when a training signal of a fixed level is transmitted to train the echo canceller, the echo of the training signal is monitored. That is, since the training signal is at a constant level (see FIG. 8), the echo level of the training signal is also substantially constant (see FIG. 9). Monitoring is performed, and when the echo of the training signal exceeds the transmission level of the training signal, it is determined that training has failed due to double talk. Then, if it is determined from this measurement that double talk is present, the apparatus is reset and retraining is performed.

As described above, in the present embodiment, the occurrence of double talk which most affects the speech quality is detected and retraining is performed, so that the speech quality can be improved. In the above-described embodiment, the level value for recognizing the occurrence of double talk is adjusted to the transmission level of the training signal. However, by setting this value to a lower level, the communication quality can be further improved. It does not matter.

[0028]

As is apparent from the above description, according to the present invention, an input exceeding the reference value is input to the echo amount calculating unit as in the case where a peak level value exceeding the training signal is input. Accordingly, the occurrence of the double talk state during training can be reliably detected by the level detection unit and the comparison unit.

[Brief description of the drawings]

FIG. 1 is a diagram showing a main configuration of an echo canceller according to the present invention.

FIG. 2 is a flowchart for explaining main operation processing of the present invention.

FIG. 3 is a flowchart illustrating a level monitoring process in FIG. 2;

FIG. 4 is a diagram showing a connection state of a conventional in-band ringer.

FIG. 5 is a flowchart for explaining main operation processing of the conventional example.

FIG. 6 is a flowchart illustrating an initial setting process in FIG. 5;

FIG. 7 is a flowchart for explaining a normal operation process in FIG. 5;

FIG. 8 is a waveform chart showing an example of a training signal.

FIG. 9 is a waveform chart showing an example of an echo level of a training signal.

FIG. 10 is a waveform chart showing a double talk state in which the amount of echo cancellation is exceeded.

[Explanation of symbols]

REFERENCE SIGNS LIST 1 in-band ringer 2 hybrid transformer 3 level monitoring unit 4 echo canceller (echo amount calculation unit, echo removal unit) 5 signal processing unit (training signal output unit) 6 signal smoothing unit (level detection unit) 7 comparison unit

 ──────────────────────────────────────────────────の Continued on the front page (72) Inventor Noriaki Nakanishi 2-1-1 Kotani, Samukawa-cho, Koza-gun, Kanagawa Prefecture Toyo Tsushinki Co., Ltd.

Claims (3)

[Claims] 1. A training signal output unit for outputting a training signal of a certain level, an echo amount calculation unit for calculating an echo amount to be removed based on a training signal output from the training signal output unit, and an echo amount calculation unit An echo removing unit that removes an echo component in a signal line based on the echo amount calculated by the above, a level detecting unit that detects a peak level value of an input signal to the echo amount calculating unit, and a peak level detected by the level detecting unit A comparison unit that compares the value with a preset reference value. If the comparison unit determines that the peak level value of the input signal exceeds the reference value, the echo amount calculation unit calculates the echo amount again. A double talk monitoring device for an echo canceller. 2. The double talk monitoring apparatus for an echo canceller according to claim 1, wherein the reference value in the comparing section is set to the same value as the peak value of the training signal output from the training signal output section. 3. The reference value in the comparing section is set to a value between a peak value of a training signal output from a training signal output section and a peak value of an echo component in the training signal. Item 2. A double talk monitoring device for an echo canceller according to item 1.